WHAT IS ARTIFICIAL INTELLIGENCE?
Artificial intelligence (“AI”) is a technical field of computer science that includes machine learning, natural language processing, speech processing, expert systems, robotics and machine vision. The term “artificial intelligence” is sometimes challenged in favor of machine intelligence or machine learning.
Machine learning automates decision making using programming rules and in some cases training data sets. Human subject matter experts can provide feedback on results as part of a training process. Machine learning can adapt its programming based on the training process and feedback. The data can be represented by various graph and network structures. For example, an artificial neural network or neural net is a system designed to process information by simulating the framework of biological brains. Deep learning involves abstract representations of data to optimize the machine learning process. Supervised learning uses labelled training data examples to infer functions that can be used for processing new data. A computer can predict or ”guess” the meaning of new data based on the training data set, graph and network structures, and feedback. Reinforcement learning involves rules to control software action in an environment to maximize a reward. Reinforcement learning may not need training data examples with labelled data sets.
Expert systems can solve scheduling, optimization, and planning problems. Speech processing includes conversion between speech (audio) and text. Natural language processing derives meaning, context, or sentiment in textual data using grammars and graph structures. Machine vision can detect patterns in visual content for object tracking, audio and face recognition, for example. Robotics involves the use of biologics systems to automate and mechanically control machine movements.
INTELLECTUAL PROPERTY FOR AI
Technology enterprises and start-ups are both competing and constructively working together to develop and deploy AI products and services. Companies and research institutions should clearly define and protect their intellectual property with registrations and documentation, especially when working with multiple third parties. A company may then control use of its IP rights, including permitted use under licensing and collaborative arrangements.
Copyright automatically extends to computer code, visual interface features, audio, video guides, application programming interface (API) structure and other works. Computer code may cover particulars such as source code, pseudo code, machine code and purpose-built hardware or firmware. Copyright is an important intellectual property asset particularly if the program design provides computational and usability efficiencies. Ownership and confidentiality of the copyright should clearly be set out in a written agreement.
Companies may also benefit from placing digital locks on their products and services for security. Circumvention of digital locks is an offence in some jurisdictions and may provide relief against unauthorized parties. Companies should have policies for developers incorporating third party copyright, even if inadvertently, as it may impact ownership of the technology and freedom to operate. Employees or a contracted developer, for example, may incorporate third-party source code without authorization which may impact ownership.
AI systems involve large data sets. These data sets and algorithms are important IP assets for the company. Contractual terms with end users and third parties should clearly specify permitted use. Some jurisdictions provide specialized IP protection for database rights.
Brands may include a word mark, logo or icon protected as registered or unregistered trade-marks, the latter of which can prevent competitors from unlawfully passing off on or diluting the goodwill of a brand. Companies can develop their brands with quality customer service and trust to establish goodwill in their brand with customers and the general public. A strong brand helps AI companies differentiate their products and services from competitors. AI technology and algorithmic accountability can help a company develop good will for their brand. Companies are often stewards of important data, assets and documentation, a reputable brand may be of paramount importance to customers.
Trade secrets are common law rights that provide protection over secret business information, and may protect material such as confidential backend server processes, code and ‘‘secret sauce.” Trade secrets require no formal registration, but companies must also take reasonable steps to keep it secret. In turn, the protected information may be protected for an unlimited period of time as long as it is kept secret and has commercial value. Misappropriation (e.g., unauthorized use) of trade secrets is regarded as unfair business practice. Trade secrets may take various forms such as customer lists, source code and technical documentation, among others.
Trade secret protection has limitations, particularly if relied on as protection for vital company assets. Trade secret rights may be difficult to establish or enforce, and enforcement may be practically ineffective against third parties who obtain the invention indirectly from an unauthorized discloser. Trade secret protection may prevent collaboration and integration with other entities in developing AI products and services. Trade secrets also do not protect against independent development of the secret innovation by third parties.
Industrial designs can be used to protect visual features of physical articles such as electronic cards, transaction machines, as well as computer interfaces, animations and icons. Design protection can be a valuable asset, especially if a given feature helps promote the distinctiveness of the brand, products and services, or increases the usability of a product.
Patents provide a mechanism to exclude others from making, using or selling the patented technology, which may help companies obtain or maintain market share, and protect research and development investments. Patents can provide a competitive advantage, and may also be used defensively as a negotiation tool. Patent publications can also be cited against subsequently filed applications to prevent grant.
A technology development strategy should consider if patent protection is available for core technology innovation. Companies should also be aware of other publications and litigations, as competitors and other players may have their own patents or pending applications. In contrast with trade secrets, granted patents may be enforced against third parties that make, use or sell the claimed invention, despite independent development. Given the quickly evolving AI market, obtaining early priority dates is important in view of the ‘‘first to file” nature of the patent system.
Generally, patents are granted worldwide for new, useful, and non-obvious inventions of patentable subject matter. Computer implemented inventions are under a greater level of scrutiny and not all AI related innovations are per se patentable. The jurisprudence determining whether technology is indeed patentable subject matter is constantly evolving. Patent offices, along with the courts, have struggled with establishing clear delineations of what is patentable and what is not patentable. Highlighting salient technical features such as technical advantages and practical implementation details can increase the likelihood of success during patent examination. The description should highlight discernible effects generated by the AI innovation or use case.
Current AI agents are winning games of Jeopardy and Go against human experts but the scientific field of cognitive computing is still relatively young. The notion that computers can simulate cognitive intelligence raises philosophical questions about whether the mind can be modelled and the ethics of programming machines for human-like decision making. Given the importance of data analytics, companies continue to invest in research and develop in AI to advance their processing and data mining capabilities.
An IP strategy for AI development and deployment will layer intellectual property rights to protect different aspects of the innovation. Companies can clearly define and protect their intellectual property with registrations and documentation. Clear agreements on intellectual property rights should be established between third parties to manage risk.
Maya Medeiros is a lawyer, patent agent, and trade-mark agent at Norton Rose Fulbright LLP Canada (Toronto). Maya Medeiros’ practice focuses on the creation and management of intellectual property assets in Canada, the United States and around the world.